AI infrastructure investor. Power, cooling, memory, networking: the bottlenecks. MA Economics | UK senior treasury analyst | Public track record since 2025
Hey Gaetano, I like others have noticed a decline in the content being pushed on here.
Reactive commentary and engagement bait slop seem to be at the forefront.
My thesis is clear and built around the bottlenecks severely constrained in the AI buildout; memory, cooling, power and connectivity.
I do not spam multiple tickers daily but instead deepen the thesis on those I am heavily invested in.
Can't wait for $META to smash earnings out of the park tonightโฆ
Forget EPS. The capex guide is what matters.
Go on Zuck, give the AI trade the confidence booster it needs.
CAPEX HIGHER
Can't wait for $META to smash earnings out of the park tonightโฆ
Forget EPS. The capex guide is what matters.
Go on Zuck, give the AI trade the confidence booster it needs.
CAPEX HIGHER
Dates to note down for the next 5 weeks:
Week 1
Oct 21 | $TSLA, $META (after close)
Week 2
Oct 28 | $MSFT, $GOOGL
Oct 29 | $AMZN, $NBIS (est.)
Week 3
Nov 2 | $AAPL
Nov 3 | US midterm elections
Nov 6 | $IREN (est.)
Week 4
Quiet week
Week 5
Nov 18 | $NVDA
Save this for later.
Agents used 34x more tokens than in February. (OpenRouter)
More tokens = more compute.
> Compute needs memory: $MU
> Agents need CPUs to orchestrate: $AMD / $INTC
> All of it needs power: $IREN
Useful AI creates demand for more intelligence. Every bottleneck tightens.